A Comparison of Dynamic Reposing and Tangent Distance for Drug Activity Prediction